02 · Grow it
Product growth for a UK GLP-1 provider
Postgres · funnel instrumentation · experiment design · product mockups
At a glance
- Problem
- A regulated product can’t be shared socially, so growth has to come from 1:1 loops, content and the customer experience, with as little friction as possible.
- Build
- A growth system for the product: a 1:1 referral flow that captures an identifying contact, funnel instrumentation that catches breaks, claim resolution, and personalised content that solves for adherence and retention.
- Result
- Smoother referrals, better customer experience, fewer support tickets, with the infrastructure to retain customers for longer so they refer more.
How it works
- Start and end
- Friend's journey
- Claim created
- Out of the run
Key decisions
A UK GLP-1 provider wants to take its 30-day K factor from 0.0835 to 0.30.
I started with the target metric. Their own numbers say the referral does not happen at checkout: 1.21% share within 1000 seconds of paying, first referral at 26.8 days. The nature of the product could explain why: the starting dose is held for four weeks, and visible weight loss usually comes after that. A 30-day K is measuring patients before the medicine has done anything worth talking about, so it is worth interrogating whether a discount on the next order is enough incentive to push customers to refer before they may be ready.
Five builds:
- Claim-first contact capture.The friend claims the discount with a phone number or email as step one, so the referral survives a device switch, a refused cookie banner and an abandoned signup. Review analysis showed lost claims are a live complaint today.
- Funnel instrumentation: 9 events.3 for the sender, 6 for the friend, joined by 4 ids that hand off in sequence. Helps determine where the funnel breaks.
- Referral resolution in Postgres.When a referred friend signs up, the system decides whether they are new or already in the database. Email and phone are unique, so it matches on the four remaining fields: first name, last name, address line, and postcode. 12 of the 16 combinations resolve on their own; 4 go to a human to check against the ID record. Claims go through cleanly, which cuts support tickets.
- Content engine.The 13 structured data points collected at onboarding drive short, clinician-approved pieces delivered through an app.
- Freemium app, 4 tabs.Dose reminders and refill prompts, a weekly content piece, weight and habit tracking, and a referral tab that sends a friend a GLP-1 primer along with the link. It gives a customer a reason to remember the product between deliveries.
Work with me
What would most help your business grow?
Thirty minutes, on video. Pick a time that suits you.
If the times do not load, open the booking page in a new tab.